OpenAI is reportedly working on a new AI-powered music generation tool capable of creating songs and instrumental tracks from text and audio prompts, according to a report by The Information. The project is said to involve collaboration with students from The Juilliard School, who are helping the company produce and annotate musical training data.
The tool could allow users to generate music to accompany videos or even create guitar backing tracks for vocal recordings, sources familiar with the project told The Information. The system would function similarly to other generative AI models, taking short written or spoken instructions and turning them into audio compositions.
It’s not yet clear how far OpenAI has progressed or whether the tool will launch as a standalone product or integrate with ChatGPT and Sora, its video generation app.
One source said OpenAI is working with Juilliard students to annotate sheet music and musical scores, a process that could help train the AI to recognize structure, rhythm, and harmony in complex compositions. The collaboration suggests OpenAI is aiming for high-quality, musically literate training data rather than relying solely on scraped online audio.
While OpenAI has previously experimented with music-generating models, those projects predate the release of ChatGPT. In recent years, the company’s audio research has focused on speech synthesis and speech recognition models instead.
The growing interest in generative music aligns with efforts by competitors such as Google, Suno, and ElevenLabs, all of which have launched tools that can produce AI-generated songs and instrumental compositions.
Featured image credits: da-kuk/Getty Images
For more stories like it, click the +Follow button at the top of this page to follow us.